Main Page Map View
Friends List Page
Settings Page

Beepd 🐝

Inspiration

Living in Vancouver, I noticed a peculiar paradox: we're surrounded by friends, yet somehow always missing each other. The "Vancouver cold" isn't just about the weather; it's about people being distant despite being physically close. How many times have you grabbed coffee alone, only to later discover a friend was two blocks away?

I wanted to turn random run-ins into deliberate connections, transforming "I wish I'd known you were nearby!" into "Let's meet up now!"

Apple's Find My iPhone showed location sharing was possible, but constant tracking felt invasive. I wanted the serendipity without sacrificing privacy a way to share my location only when I'm open to meeting up, not 24/7 surveillance of my every move.

What It Does

Beepd is a privacy-first friend discovery app that helps you connect IRL when it actually matters.

Toggle Location Sharing: Big ON/OFF switch—share your location only when you're ready to meet friends
Smart Proximity Detection: Get notified when friends enter your customizable radius (100m, 500m, or 1km)
Privacy Modes: Choose between OFF, Friends-only, or Everyone
Friend Codes: Add trusted friends via shareable codes—no phone numbers, no social graphs
Live Map View: See friends clustered on an interactive map with real-time updates
No Spam Notifications: One-time alerts when friends newly enter your radius, not constant pings

How I Built It

Frontend: React + TypeScript with Vite, using React Leaflet for interactive maps and Supercluster for efficient marker clustering. Radix UI and Tanstack Query power a smooth, responsive experience.

Backend: Cloudflare Workers with Hono framework for edge-deployed serverless APIs. Drizzle ORM with D1 (SQLite) for blazing-fast queries at the edge.

Key Technical Decisions:

Haversine distance formula for accurate proximity calculations: $$d = 2r \arcsin\left(\sqrt{\sin^2\left(\frac{\phi_2-\phi_1}{2}\right) + \cos(\phi_1)\cos(\phi_2)\sin^2\left(\frac{\lambda_2-\lambda_1}{2}\right)}\right)$$
Token-based auth with secure friend code generation for frictionless onboarding

Challenges I Ran Into

Privacy vs. Functionality: Balancing real-time location sharing with user privacy. Solved with explicit opt-in modes and temporary location storage.
Demo on a Laptop: Testing proximity features without multiple devices. Built a location simulator that lets you spoof coordinates per browser window, enabling me to demo two users side-by-side!
Map Performance: Rendering hundreds of markers crashed the browser. Implemented Supercluster for efficient clustering and viewport-based rendering.
Accurate Distance Calculations: Simple Euclidean distance doesn't work on a sphere! Implemented the Haversine formula for precise geospatial calculations.
Notification Spam Prevention: Friends shouldn't get pinged every second they're nearby. Built debouncing logic to trigger alerts only on new proximity entries.

What I Learned

Edge computing with Cloudflare Workers delivers sub-50ms API responses globally
Privacy by design isn't just a feature; it's a fundamental architectural decision
User psychology matters: A big ON/OFF toggle provides peace of mind that "incognito mode" never could
Demo-ability drives development: Building the location simulator made testing 10x easier and the project actually demostrable
Github Specify: Built this project hand-in-hand with Copilot integration with advanced planning and requirements meant build time decreased immensely

What's Next for Beepd

Calendar Integration: Find mutual free time and schedule hangouts automatically
iOS/Android Apps: Native mobile experience with background location updates
Event Mode: Create temporary location-sharing groups for concerts, conferences, or festivals
Safety Features: Panic button, trusted contacts, and time-limited sharing for added security